Expression of a Transgenic T Cell Receptor/3 Chain Enhances Collagen-induced Arthritis By Lucia Mori, HansruediLoetscher, Kiichi Kakimoto,* Horst Bluethmann, and Michael Steinmetz From the Pharmaceutical Research New Technologies, F. Hoffmann-La Roche, Ltd., 4002 Basel, Switzerland; and the *Kumamoto University Medical School, Institute for Medical Immunology, gumamoto 860, Japan
Summary SWR/J transgenic (tg) mice were generated expressing the TCR ~ chain derived from an anticollagen type II (CII) arthritogenic T cell clone. The SWR/J strain was selected because it is resistant to collagen-induced arthritis (CIA) and lacks the V/3 gene segment used by the T cell done. Expression of the tg/3 chain on all thymocytes and peripheral lymph node T cells led to a more efficient anti-CII immune response, but did not confer CIA susceptibility to SWR/J mice. Nevertheless, this tg/3 chain enhanced predisposition to CIA as (DBA/1 x SWR) F1 tg mice were more susceptible than normal F1 littermates. Our results demonstrate that the expression of the tg/3 chain contributes to CIA susceptibility, but by itself it is not sufficient to overcome CIA resistance in the SWR/J strain.
ollagen-induced arthritis (CIA) 1 develops in certain strains of mice, rats, and primates after immunization C with native type II collagen (CII) in CFA (1-3). The disease can be induced by homologous as well as heterologous CII. Cyanogen bromide digestion of chicken CII gives rise to a major immunogenic and arthritogenic fraction which is represented by the CBll peptide (4). Both antibody and T cell responses induced by heterologous CII immunization are predominantly directed to epitopes that are present in the CB11 peptide (5, 6). In mice, susceptibility to the disease is associated with the expression of particular MHC class II alleles such as Aq and A' (7, 8). Several polymorphic genes at other loci also influence disease susceptibility. The most extensively studied ones are genes encoding complement component C5, TCR chain, and Mls antigens (9-12). Several investigators have attempted to clarify the putative role of C5 in susceptibility to CIA. The results are difficult to interpret and are in part contradictory (9-11). The crucial role of T cells in CIA has been dearly documented in many studies (13). The disease can be transferred by T cells, either using cells fleshly isolated from CIIimmunized animals or using CII-specific Th cell lines and clones (14-16). Furthermore, CIA development can be prevented by injections ofmAb directed against the CD4 and 1 Abbreviations used in this paper: BCG, Bacillus Calmette-Gu*rin; CIA, collagen-induced arthritis; CII, collagen type II; tg, transgenic.
381
TCR molecules (17-19). Because of the linkage to particular class II alleles, one would expect that the "pathogenic T cells" may express a rather restricted TCR repertoire. Indeed, several findings suggest that the recognition of the diseaseinducing CII epitope(s) requires particular TCR B chains. Mice with the susceptible H-2 background, but with a genomic deletion of certain TCR V/~ gene segments, are resistant to the induction of CIA. These mouse strains have a deletion encompassing the V~9, V~11, V~12, V~13, V~5, and V~8 gene families, and have been classified as VB~ mutants (20, 21). Among these, SWR/J and AU/SSJ strains, both of the H-2q haplotype, are CIA resistant (22, 23). Also, Rill S/J mice, which have an even larger deletion including Vfl6, Vf115, and Vf117 genes (Vflc mutants), are CIA resistant despite being H-2 r (24). The correlation between resistance to the disease and deletion of certain Vfl genes suggests that these genes play a role in the pathogenesis of CIA. Although this hypothesis is supported by gene complementation and segregation experiments, the Vfl involvement in the control of CIA remains controversial (9-11, 22). Recent studies on CIA in Mls-1 congenic mouse strains suggest that expression of a particular endogenous superantigen, Mls-P, results in decreased CIA incidence (12). Endogenous superantigens are characterized by their ability to be recognized by subsets of T lymphocytes bearing particular Vfl chains. Expression of the superantigen in the thymus of mice with a permissive H-2, results in clonal deletion of the T cells reactive to this superantigen. The consequence of the elimination of these autoreactive T cells is tolerance
J. Exp. Med. 9 The Rockefeller University Press 9 0022-1007/92/08/0381/08 $2.00 Volume 176 August 1992 381-388
to self superantigens (25, 26). With respect to CIA, the expression of Mls antigens may cause important ddetions in the V~ peripheral repertoire affecting T cells that are possibly important in the pathogenesis of the disease. The TCR V~ involvement in the pathogenesis of arthritic disease was also proposed upon analysis of TCR genes expressed in T lymphocytes present in the synovia of rheumatoid arthritis patients. In some (27-29), but not all of these studies (30-32), an expansion of oligoclonal T calls, or a restricted use of certain V~ chains (33, 34) was observed. In the present study we have analyzed the role of TCR V~ genes in CIA using a different approach, namely by generating SWR/J mice that are transgenic (tg) for a TCR/3 gene derived from an anti-CII arthritogenic T cell done. The particular T cell clone that we have used transfers the disease and expresses a V~ region normally absent in the genome of the CIA-resistant strain, SWR/J (22). The tg SWR mice express the tg ~ chain on all T lymphocytes, and after CII immunization, show increased T cell-dependent anti-CII antibody responses, but do not develop CIA. However, introduction of the transgene into susceptible (DBA/1 x SWR/J)F1 mice clearly increases CIA predisposition.
Materiak and Methods Mice. DBA/1J,SWR/J, and MORO/Ibm mice were obtained from BiologicalResearch Laboratories, (F~illinsdorf,Switzerland). Transgenic mice were alsobred and maintained in the mouse colony at Hoffraann-LaRoche (Basal, Switzerland). F1 crosses were made between SWR/J male mice heterozygous for the B transgene and DBA/1J female mice. For immunofluorescence analyses, unimmunized mice were killed at 4-8 wk of age and CII immunized at day 42 after immunization. T Cells and Culture Conditions. Ceils were grown in IMDM (Gibeo, Paisley,UK) supplementedwith 10% heat-inactivatedFCS (Flow, UK), 2 mM 1,glutamine, 1 mM Na pyruvate (Gibco), 5 x 10-s M 2-ME (Sigma Chemical Co., St. Louis, MO), 50 U/m1 penicillin, 50 ~g/ml streptomycin, 100/~g/ml kanamycin(Gibco), and incubated in a humidified incubator with 5% CO2. The antiCII T cell clone K-102 was derived from the anti-CII T cell line previouslydescribedby Kakimoto et al. (15). The clone was characterized with respect to antigen specificityand shown to passively transfer arthritis in DBA/1 recipients (K. Kakimoto, unpublished results). K-102 cells (2 x 10Vml)were stimulated with 40/~g/ml human CII in the presence of 107/ml DBA/1 irradiated (3,000 tad) spleen ceils. Between stimulations, cells were maintained in culture medium supplementedwith 10% supernatant of Con A-stimulated rat spleen cells and 2 rag/m1 methyl o~-o-mannopyranoside (Sigma Chemical Co.) (35). 10 d after CII stimulation, blasts were purified over a ficoll gradient (Cederlane, Hornby, Canada) and fused with a subclone of BW-1100.129.237 ol-B- lymphoma cells (36). T cell hybridomas were obtained after HAT selection, subcloned, and analyzed for antigen specificity, and CD3 and CD4 expression. Collagens and Preparationsof cyanogen bromide(CNBr) Digests of CIL Human and mouse type II collagenswere purifiedfrom costal and sternal cartilages according to the method describedby Miller et al. (37). The purity of CII was assessedby SDS-PAGE. Bovine CII was a kind gift from M. J. Mullqeen and D. Bradshaw (Roche Products, Welwyn Garden City, UK). Chicken CII was purchased
382
from Genzyme Corp. (Boston, MA), bovine type I, human type III, and human type IV collagens were all purchased from Sigma Chemical Co. Bovine and chicken CII were digested with CNBr in 70% formic acid at 37~ After desalting on a Sephadex G10 column in 0.1 M aceticacid, the digest was applied to a S-Sepharose column (Pharmacia, Uppsala, Sweden) equilibrated in 20 mM citrate, 20 mM NaC1 pH 3.8 buffer, and eluted with a linear NaC1 gradient (20-500 mM). The temperature of the column was maintained at 45~ Fractions containing stimulator), activity in the functional assay(seebelow) were pooled, lyophilized,and subjected to gel filtration chromatographyin PBS on a TSK 3000SW preparative column (LKB-Pharmacia,Gaithersburg, MD). The purity of the ehted peptides was checkedby SDS-PAGEaccordingto Laemmli (38), using a mini-gel system (Bio-Rad Laboratories, Richmond, CA). Partial amino acid sequences were determined on a protein sequencer (475 A; Applied Biosystems, Inc., Foster City, CA). Antigen-specific11.,2ReleaseAssay. The T cell hybridoma (5 x 104 cells/well) was tested with a titration of collagens or CII CNBr peptides in the presence of DBA/1 or SWR/J spleen cells (5 x 10s cells/well) in flat-bottomed 96-well plates (Costar, Cambridge, MA). After 20-24 h of incubation, cell-free supernatants were collectedand tested for I1,2 content using the CTL1,2 proliferation assay. CTL1,2 were plated at 5 x 103/well, and after 2024 h of culture were pulsed for an additional 6 h with 1/~Ci/well of [3H]thymidine (Amersham International, Amersham, Bucks UK). Incorporated radioactivity was measured with a liquid scintillation B counter (Betaplate; LKB-Wallac,Turku, Finland) and is expressed as mean cpm of triplicate cultures. Induction and Evaluation of CIA. Mice (8-10 wk) were immunized by intradermal injection in the back at the base of the tail with 200/~g of bovine or chicken CII in CFA (Difco Laboratories, Detroit, Michigan). An intraperitoneal injection of Mycobacterium bovis Bacillus Calmette-Gu6rin (BCG, ,,o0.8-2.6 x 106 cells/mouse; Schweiz. Serum- & Impfmstitut, Bern, Switzerland) was also given to the mice at the same time. 21 d after the immunization, mice were boosted intraperitoneally with 200 #g of CII together with BCG. Arthritis was monitored daily from day 21 until day 35, weekly during the second month, and then every 2 wk thereafter. Mice were monitored for up to 3 mo, and in the case ofSWR fl tg mice, for up to 5 mo after immunization. Dates of onset of disease were recorded for individual mice. CIA was diagnosed if redness and swelling of fore- and hindpaws were observed. Severityof arthritis was evaluatedfor each paw and graded as following: grade 1, redness and swelling; grade 2, deformity; and grade 3, ankylosis.The scores for each paw were added to give an index ranging from 0 to 12 per mouse. None of the mice ever developed CIA after CFA/BCG injection. Statistical analysis was done using the X2 test (for comparing CIA incidence) and the Mann-Whitney test (for comparing day of onset and severity of CIA). TCR 3 Gene Constructand Generation of SWR ~ tg Mice. The productively rearranged TCR B gene has been isolated by PCR amplification from genomic DNA of the anti-CII T cell clone. Primers used are complementary to sequences located in the 5' flanking region of the VB12 leader sequence containing an additional SalI cloning site (5' primer), and in the intron between JB1.2 and JB1.30' primer), allowing the amplificationof a DNA region containing a BamHI site at the 3' ofJB1.2, pCR was performed for 35 cycles (94~ 40 s; 65~ 50 s; and 72~ 60 s) followed by 5 rain at 72~ using Taq DNA polymerase(Perkin-ElmerCetus, Emeryville, CA). The amplifiedfragment of ~900 bp was inserted into a plasmid vector containing TCR B enhancer and VB8.3 pro-
Collagen-inducedArthritis in TCP. 13TransgenicMice
moter sequences (C. E Gray, Hoffmann-La Roche, Basel, Switzerland). A 6.4-kb fragment spanning the region from the BamHI site located in the J/31 locus to the HindlII site at the end of the C/31 untranslated region, was further ligated into the construct at the corresponding BamHI site (see Fig. 2). This provides the entire C/31 region used by the K-102 T cell clone. After DNA sequencing to control possible PCK and ligation artifacts, the/3 construct was excised from the prokaryotic vector (pBluescript II KS- ; Stratagene, La Jolla, CA) by digestion with BssHII, separated on agarose gd, and purified by phenol/chloroform extractions followed by ethanol precipitation. Transgenic mice were generated by microinjection of the purified DNA into SWR/J fertilized oocytes that were reimplanted into the oviducts of MOP,O/Ibm foster mothers as described (39). The integration of the transgene was analyzed both by PCR and Southern blot using DNA isolated from tail biopsies of 2-3-wk-old mice. The transcription of the fulllength tg B RNA was demonstrated by Northern blot using total spleen RNA. AntibodiesandFlowCytometry. The following mAb were used: anti-CD3 e biotin-conjugated (500A2; PharMingen, San Diego, CA) or FITC-conjugated (145-2Cll; Boehringer Mannheim, Mannhelm, Germany); anti-CD4 PE-conjugated (GK1.5; Becton Dickinson & Co., Mountain View, CA); anti-CD8 FITC-conjugated (53-6.7; Becton Dickinson & Co.); anti-Thy 1.2 FITC-conjugated (30-H12; Becton Dickinson & Co.); anti-VB2 (B20.6.5, kindly provided by B. Malissen, MarseiUe-Luminy, France); anti-V~3 (KJ25a; J. Kappler, Howard Hughes Medical Institute, Denver, CO); anti-VB4 (KT4-3; K. Tomonari, Medical Research Coundl, Harrow, U.K.); anti-V~6 (44-22-1; H. Hengartner, University Hospital, Ziirich, Switzerland); anti-V~8.1, 8.2, and 8.3 (U. Staerz, National Jewish Center, Denver, CO); anti-VBll (KT11; K. Tomonari); and anti-V~17a (KJ23a, J. Kappler) were all used as cell culture supernatants. The MR11-1 hybridoma secreting anti-V/312 mAb was produced by O. Kanagawa (Washington University, St. Louis, MO), and obtained through the courtesy of H. R. MacDonald (Ludwig Institute for Cancer Research, Lausanne, Switzerland). The antibody was purified from the cell culture supernatant and biotinylated by standard procedures using N-bydroxysuccinimidobiotin (Sigma Chemical Co.). Second step reagents were: sheep anti-mouse Ig-FITC, and sheep anti-rat Ig-FITC (Silenus, Hawthorn, Australia); goat anti-mouse Ig-PE, and goat anti-rat Ig-PE (Southern Biotechnology Associates, Birmingham, AL); and streptavidin-FITC or streptavidin-conjugated tandem label of PE/Texas Red (Southern Biotechnology Associates). All incubations and washings were done at 4~ in PBS, 2% FCS, 0.02% NAN3. 20,000 or 50,000 viable cells were analyzed by FACScan| (Becton Dickinson & Co.). Dead cells were excluded from the analysis using forward and side scatter parameters and also using propidium iodide when possible. Serum Anti-CII Antibody Levels. ELISA was performed as follows: fiat-bottomed Maxi Sorp 96-well Immuno Plates (Nunc, Roskilde, Denmark) were coated overnight at 4"C with 50/~l/well of either CII or purified protein derivative from Mycobacteriumtuberculosis(PPD, Statens Seruminstitut, Copenhagen, Denmark), BSA (Sigma Chemical Co.), or goat anti-mouse Ig (Southern Biotechnology Associates), all at 10/zg/ml. Plates were saturated with 1% BSA in PBS, and after washing, sera were added at various dilutions. After five washes with PBS, 0.05% Tween 20, ~ galactosidasehbeled antibodies (anti-mouse Ig, IgM, IgG1, IgG2a, IgG2b, IgG3, Southern Biotechnology Associates)were added and incubated overnight at 4~ Positive wells were revealed with O-nitrophenyl-~n-galactopyranoside (Sigma Chemical Co.) and ODium was determined using a microplate reader (Anthos Labtec, Basel, Switzerland). 383
Mori et al.
Results
The T Cell Hybridoma BL17 Recognizes the CBlI Fragment of CII. A T cdl hybridoma, BL17, was derived from the fusion of the arthritogenic anti-CII T cell done K-102 and the BW-1100.129.237 c~-~- T cell lymphoma (see Materials and Methods). This hybridoma expresses high levels of CD3 and CD4. Stimulation by immobilized anti-c~/~ TCR or antiCD3 antibodies leads to Ib2 release and apoptosis (data not shown). Like the parental clone, the hybridoma produces Ib2 in response to bovine, chicken, human, and mouse CII presented by DBA/1 APC (data not shown). To further characterize the antigen specificity, bovine and chicken CII were digested with CNBr, and the peptides were fractionated by ion exchange and gel filtration chromatography. The fractions were tested for their ability to induce Ib2 release in the presence of APC. Stimulatory activity could be attributed to a peptide migrating with an apparent molecular mass of about 36 kD on SDS-PAGE (data not shown). Partial amino acid sequences of the stimuhtory peptides from both bovine and chicken CII showed that they were identical with the previously described CBll peptide (40), a 279-amino acid-long peptide that is the main CII fragment recognized by both antibodies and T ceils in mice with CIA.
$ WR/JSpleen Cells Can Processand PresentCII and its CB11 Fragment to the T Cell Hybridoma. S W R / J mice are not susceptible to CIA induction, perhaps because of the deletion of some V~ gene segments. An alternative explanation could be that although they carry susceptibility-associated H-2 genes, their APC might be unable to present arthritogenic CII epitopes. We have excluded this possibility by showing that spleen cells from S W R / J mice are able to process and present either bovine, chicken and human CII (Fig. 1 a), or bovine and chicken CBll (Fig. 1 b) to the BL17 hybridoma. Thus, the resistance of S W R / J to CIA is not due to a defect in antigen presentation. Generation of TCR B TransgenicMice. We have attempted to sensitize the S W R mice to CIA by introducing a transgene encoding the BL17 TCR ~ chain. This T C R B chain was shown to be encoded by V/512-D~l.l-J~l.1 and C~/1 (L. Mori, unpublished results). The productively rearranged
80
~
10
40 ZO
.01
.1
1
I0
1~
IOOQ .OOG1.001 .01
.1
1
10
pglml
Figure 1. Antigenpresentationby SWR/J spleencells. BL17cellswere stimulated by CII or its CBll fragment in the presenceof SWR/J APC. The levelsoflb2 secretedweremeasuredusingthe CTLL-2assay.(a) Shows the stimulation with bovine CII (11), human CII (0), and chicken CII (A); (b) with bovine CB11 ([~) and chicken CB11 (A). Results are expressed as mean cpm of [3H]thymidineincorporation of triplicate cultures.
LVI312DI31.1J]31.1 5'
El3
PVI~8"31 ~
131
a SWR/J
Figure 2. TCR ~ construct used for the generation of transgenic mice. The construct is composed of a PCR-amplified fragment that contains the leader sequence of VB12 (L), V~12, DBI.1, and JBI.1 (nucleotide sequence is available upon request). This fragment is inserted 3' to the/3 locus enhancer (E/3) and promoter (P) of V/iS.3, and 5' to the entire CB1 lOCUS.
~_
TCK B gene was isolated by PCR amplification of genomic DNA isolated from the anti-CII T cell clone K-102, and inserted into a plasmid vector containing the TCR/3 enhancer and VBS.3 promoter sequences. A fragment encompassing the entire C/51 locus was further inserted at the 3' end of the V/5 gene (Fig. 2). This TCK/5 construct, free from v e c t o r sequences, was used for the generation of tg mice. Two tg mouse lines were obtained in the SWK/J strain. All experiments described in this study were carried out with one of these lines (SWR-/SL tg) which integrated about 10 copies of the transgene into the genome and expressed very high levels of tg/5 chain mRNA in the spleen (data not shown).
Expression of the TCR /8 Transgenein the Absence of Endogenous B Genes. The expression of the TCR/5 transgene was tested using a mAb specific for the mouse V/512 polypeptide. Thymocytes and lymph node cells from tg mice and non-tg littermates were stained with mAb specific for CD4, CD8, and V/512, and evaluated by three-color FACS| analysis. In the thymus, V/512+ calls were undetectable in SWK mice, as expected, but were present in SWR-BL tg animals (Fig. 3, a and d). The number of V/512+ and CD3 * ceils
r
b 36
3
SWR/J 5d
e
~s
~
f
20
~3
[ .*
'" 4
SWR [3L tg
V~12
CD3
CD8
Figure 3. Thymocyte surface expression of the transgenic/3 chain in SWR-BL tg mice. Thymocytcs from SWK-BL tg and non-tg, sex-matched, littermate mice were stained with anti-CD8 (FITC-conjugated), anti-CD4 (PE-conjugated) and anti-V/~12 or anti-CD3 (biotin-conjugated), followed by labeling with streptavidin-tandem (see Materials and Methods). (a and d) Histograms of the Vf112 staining; (b and e) Histograms of the CD3 staining; (c andf ) Contour plots for CD4 and CD8 stainings ofceUs gated as CD3 + . Numbers represent the percentage of cells in each region. The difference in staining of the TCR. k'~ population, using anti-Vf112 or antiCD3, has to be attributed to different ai~nities of the two mAb. 384
iI
e L~ r~
f
t
d SWR JSLtg
V~312
a
c
3' lKb
Tg ~
b
CD3
CD8
Figure 4. Lymph node T cell surface expression of the transgenic/~ chain in SWR-BL tg mice. Total lymph node cells from SWR-BL tg and nontg, sex-matched, littermate mice were analyzed by three-color staining as in Fig. 3. (a and d) Histograms of the V~12 staining; (b and e) Histograms of the CD3 staining; (c and f ) Contour plots for CD4 and CD8 stainings of cells gated as CD3 +. Numbers represent the percentage of cells in each region.
was similar, thus indicating that every T cell expresses the transgene. Although the total number of CD3 + thymocytes was similar in normal and tg animals, there were differences in the percentages of mature cells. Tg mice had a lower number of CD3 hi thymocytes (3% vs 12% in normal animals) and this was compensated by a higher number of immature, CDY ~ cells (49% vs 45%; Fig. 3, b and e) and CD3+CD4+CD8 + cells (63% vs 53%; Fig. 3, c and f). In lymph nodes of tg mice all CD3 + cells were V/512+ and were reduced in number (52% vs 74% of total lymph node cells; Fig. 4, a, b, d, and e), but the ratio of CD4 to CD8 cells was as in normal mice (Fig. 4, c and f). We also used other VO-specific mAb (anti-VB 2, 3, 4, 6, and 17a) to show the lack of expression of endogenous/3 genes. Indeed, none of these reagents stained either thymocytes or lymph node cells in SWR-/SL tg animals (data not shown). These results show that the tg/~ chain is expressed on the surface of T cells and that it prevents the expression of the endogenous TCR/3 chains. Furthermore, expression of this chain also affects the number of mature cells, without altering the relative percentages of CD4 + and CD8 + subsets.
SWR-~L tg Mice Mount Strong Anti-CII Antibody Responses. The antibody response of mice to CII is strictly T cell dependent (41). SWK/3 tg and non-tg littermates, as weU as DBA/1 mice were immunized with bovine CII in CFA. At different time points after immunization, mice were bled and the anti-CII antibodies were analyzed by ELISA. It is interesting that the SWR-BL tg mice produce more antiCII antibodies (day 7 titer, 8; day 14 titer, 50; and day 21 titer, 400) than DBA/1 mice (day 14 titer, 15, and day 21 titer, 200) and SWR non-tg littermates (day 14 titer, 3, and day 21 titer, 25; Fig. 5). Furthermore, their response developed earlier as compared with the other groups. Already at day 7 after immunization, SWR-/SL tg mice had detectable titers of anti-CII antibodies of the 3/1, 3'2b, and 3'2a isotypes. The anti-CII antibody titer in SWR-/SL tg mice was always
Collagen-induced Arthritis in TCR/3 Transgenic Mice
1.2
Table 1. SWR-~L tg Mice are Resistant to CIA
Day 7
Mice*
0.8
Arthritic/Total
%
0.4
Experiment 1 0.0
0.8
~2 0
0.4"
0.0
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
1.2 0.8'
0.4'
0.0
.
10 0 101
.
.
.
.
10 2 10 3 10 4 10 5
1 : serum dilution
Figure 5. Increasein the total anti-CII antibodytiter in the primary response of SWK-BL tg mice. Titration of the total antibovineCII Ig presentin the serum of mice immunized with bovine CII in CFA (opensymbols)or only injectedwith CFA(cbsedsymbols) on differentdaysafter immunization: SWR-/~L tg ([], I); SWR/J (A, a); DBA/1 (O, O). Each curve represents the values obtained from the pooled sera of three mice. Results are expressed in OD~nm. Control ELISAwith BSA-coated plates gave background values ~0.172.
higher than in S W R non-tg littciraates for all antibody dasses and at all time points (data not shown). Thus, in the SWR-BL tg mice there seems to be an early switch to all Ig isotypes and in particular to ~/2a that predominates in the anti-CII antibody response during CIA (41). This finding, together with the fact that the antibody titer is higher in the tg mice, suggests that these mice may have an increased number of CII-specific Th cells. SWR-~L tg Mice Are Resistant to CIA. SWR-/~L tg mice were tested for the development of CIA, to determine whether the expression ofa Vf112 bearing TCR/3 chain on all T cells, and the increased production of anti-CII antibodies observed in these mice were sufficient to make them susceptible to the disease. A total of 31 SWIL-/~ tg mice and 27 S W R non-tg littermates were immunized with bovine CII in CFA in three separate experiments. 21 d after the immunization, mice were boosted with bovine CII and analyzed for the development
Table
Incidence
2.
Experiment 2
Experiment 3
SWK-BL tg SWR/J DBA/1 SWIL-/~Ltg SWR/J DBA/1 SWR-/3L tg SWR/J DBA/1
0/12 0/14 7/11 0/13 0/8 6/7 0/6 0/5 4/5
0 0 64 0 0 86 0 0 80
* Male and female mice were used.
of arthritis. As shown in Table 1, none of the S W K developed clinical signs of CIA regardless of the presence of the transgene, even 5 mo after the first injection. In contrast, 64-86% of the DBA/1 mice showed clinical signs of CIA, starting from day 21 to day 33 after CII immunization. Expression of the tg ~ Chain in Susceptible Mice Leads to In. creased Incidence and Severity of CIA. Susceptibility to CIA is a dominant trait since (DBA/1 x SWK) F1 mice develop CIA (10, 11). Both tg and non-tg F1 littermates, derived from crossing S W K mice heterozygous for the/~ tg with DBA/1, were immunized with CII and observed for the development of CIA. Arthritis developed in 38% of (DBA/1 x SWK) F1 mice and in 89% of the F1/3L tg mice in a total of three independent experiments (Table 2). The disease induced in F1 BL tg mice was more severe (median arthritic score 8 vs 4) and developed 2 d earlier than in F1 nontg littermates. V/312 was expressed in the tg mice by all T cells present in lymph nodes draining the affected limbs, as well as in the corresponding lymph nodes of the tg mice without clinical CIA (Fig. 6). FACS| analysis using other V/3-specific mAb (anti-V/3 2, 3, 4, 6, 8, 11, and 17a), showed that none of the F1/~L tg animals also expressed endogenous/3 genes (data not shown).
CIA in (DBA/1 x SWR-ffL tg) F~ Mice
Mice
Arthritic/Total
Incidence*
Day of onset**
Severity* **
26 (21-29) 28 (28-36)
8 (12-3)
% F1/3L tg littermates Ft non-tg littermates
25/28 8/21
* p
Summary SWR/J transgenic (tg) mice were generated expressing the TCR ~ chain derived from an anticollagen type II (CII) arthritogenic T cell clone. The SWR/J strain was selected because it is resistant to collagen-induced arthritis (CIA) and lacks the V/3 gene segment used by the T cell done. Expression of the tg/3 chain on all thymocytes and peripheral lymph node T cells led to a more efficient anti-CII immune response, but did not confer CIA susceptibility to SWR/J mice. Nevertheless, this tg/3 chain enhanced predisposition to CIA as (DBA/1 x SWR) F1 tg mice were more susceptible than normal F1 littermates. Our results demonstrate that the expression of the tg/3 chain contributes to CIA susceptibility, but by itself it is not sufficient to overcome CIA resistance in the SWR/J strain.
ollagen-induced arthritis (CIA) 1 develops in certain strains of mice, rats, and primates after immunization C with native type II collagen (CII) in CFA (1-3). The disease can be induced by homologous as well as heterologous CII. Cyanogen bromide digestion of chicken CII gives rise to a major immunogenic and arthritogenic fraction which is represented by the CBll peptide (4). Both antibody and T cell responses induced by heterologous CII immunization are predominantly directed to epitopes that are present in the CB11 peptide (5, 6). In mice, susceptibility to the disease is associated with the expression of particular MHC class II alleles such as Aq and A' (7, 8). Several polymorphic genes at other loci also influence disease susceptibility. The most extensively studied ones are genes encoding complement component C5, TCR chain, and Mls antigens (9-12). Several investigators have attempted to clarify the putative role of C5 in susceptibility to CIA. The results are difficult to interpret and are in part contradictory (9-11). The crucial role of T cells in CIA has been dearly documented in many studies (13). The disease can be transferred by T cells, either using cells fleshly isolated from CIIimmunized animals or using CII-specific Th cell lines and clones (14-16). Furthermore, CIA development can be prevented by injections ofmAb directed against the CD4 and 1 Abbreviations used in this paper: BCG, Bacillus Calmette-Gu*rin; CIA, collagen-induced arthritis; CII, collagen type II; tg, transgenic.
381
TCR molecules (17-19). Because of the linkage to particular class II alleles, one would expect that the "pathogenic T cells" may express a rather restricted TCR repertoire. Indeed, several findings suggest that the recognition of the diseaseinducing CII epitope(s) requires particular TCR B chains. Mice with the susceptible H-2 background, but with a genomic deletion of certain TCR V/~ gene segments, are resistant to the induction of CIA. These mouse strains have a deletion encompassing the V~9, V~11, V~12, V~13, V~5, and V~8 gene families, and have been classified as VB~ mutants (20, 21). Among these, SWR/J and AU/SSJ strains, both of the H-2q haplotype, are CIA resistant (22, 23). Also, Rill S/J mice, which have an even larger deletion including Vfl6, Vf115, and Vf117 genes (Vflc mutants), are CIA resistant despite being H-2 r (24). The correlation between resistance to the disease and deletion of certain Vfl genes suggests that these genes play a role in the pathogenesis of CIA. Although this hypothesis is supported by gene complementation and segregation experiments, the Vfl involvement in the control of CIA remains controversial (9-11, 22). Recent studies on CIA in Mls-1 congenic mouse strains suggest that expression of a particular endogenous superantigen, Mls-P, results in decreased CIA incidence (12). Endogenous superantigens are characterized by their ability to be recognized by subsets of T lymphocytes bearing particular Vfl chains. Expression of the superantigen in the thymus of mice with a permissive H-2, results in clonal deletion of the T cells reactive to this superantigen. The consequence of the elimination of these autoreactive T cells is tolerance
J. Exp. Med. 9 The Rockefeller University Press 9 0022-1007/92/08/0381/08 $2.00 Volume 176 August 1992 381-388
to self superantigens (25, 26). With respect to CIA, the expression of Mls antigens may cause important ddetions in the V~ peripheral repertoire affecting T cells that are possibly important in the pathogenesis of the disease. The TCR V~ involvement in the pathogenesis of arthritic disease was also proposed upon analysis of TCR genes expressed in T lymphocytes present in the synovia of rheumatoid arthritis patients. In some (27-29), but not all of these studies (30-32), an expansion of oligoclonal T calls, or a restricted use of certain V~ chains (33, 34) was observed. In the present study we have analyzed the role of TCR V~ genes in CIA using a different approach, namely by generating SWR/J mice that are transgenic (tg) for a TCR/3 gene derived from an anti-CII arthritogenic T cell done. The particular T cell clone that we have used transfers the disease and expresses a V~ region normally absent in the genome of the CIA-resistant strain, SWR/J (22). The tg SWR mice express the tg ~ chain on all T lymphocytes, and after CII immunization, show increased T cell-dependent anti-CII antibody responses, but do not develop CIA. However, introduction of the transgene into susceptible (DBA/1 x SWR/J)F1 mice clearly increases CIA predisposition.
Materiak and Methods Mice. DBA/1J,SWR/J, and MORO/Ibm mice were obtained from BiologicalResearch Laboratories, (F~illinsdorf,Switzerland). Transgenic mice were alsobred and maintained in the mouse colony at Hoffraann-LaRoche (Basal, Switzerland). F1 crosses were made between SWR/J male mice heterozygous for the B transgene and DBA/1J female mice. For immunofluorescence analyses, unimmunized mice were killed at 4-8 wk of age and CII immunized at day 42 after immunization. T Cells and Culture Conditions. Ceils were grown in IMDM (Gibeo, Paisley,UK) supplementedwith 10% heat-inactivatedFCS (Flow, UK), 2 mM 1,glutamine, 1 mM Na pyruvate (Gibco), 5 x 10-s M 2-ME (Sigma Chemical Co., St. Louis, MO), 50 U/m1 penicillin, 50 ~g/ml streptomycin, 100/~g/ml kanamycin(Gibco), and incubated in a humidified incubator with 5% CO2. The antiCII T cell clone K-102 was derived from the anti-CII T cell line previouslydescribedby Kakimoto et al. (15). The clone was characterized with respect to antigen specificityand shown to passively transfer arthritis in DBA/1 recipients (K. Kakimoto, unpublished results). K-102 cells (2 x 10Vml)were stimulated with 40/~g/ml human CII in the presence of 107/ml DBA/1 irradiated (3,000 tad) spleen ceils. Between stimulations, cells were maintained in culture medium supplementedwith 10% supernatant of Con A-stimulated rat spleen cells and 2 rag/m1 methyl o~-o-mannopyranoside (Sigma Chemical Co.) (35). 10 d after CII stimulation, blasts were purified over a ficoll gradient (Cederlane, Hornby, Canada) and fused with a subclone of BW-1100.129.237 ol-B- lymphoma cells (36). T cell hybridomas were obtained after HAT selection, subcloned, and analyzed for antigen specificity, and CD3 and CD4 expression. Collagens and Preparationsof cyanogen bromide(CNBr) Digests of CIL Human and mouse type II collagenswere purifiedfrom costal and sternal cartilages according to the method describedby Miller et al. (37). The purity of CII was assessedby SDS-PAGE. Bovine CII was a kind gift from M. J. Mullqeen and D. Bradshaw (Roche Products, Welwyn Garden City, UK). Chicken CII was purchased
382
from Genzyme Corp. (Boston, MA), bovine type I, human type III, and human type IV collagens were all purchased from Sigma Chemical Co. Bovine and chicken CII were digested with CNBr in 70% formic acid at 37~ After desalting on a Sephadex G10 column in 0.1 M aceticacid, the digest was applied to a S-Sepharose column (Pharmacia, Uppsala, Sweden) equilibrated in 20 mM citrate, 20 mM NaC1 pH 3.8 buffer, and eluted with a linear NaC1 gradient (20-500 mM). The temperature of the column was maintained at 45~ Fractions containing stimulator), activity in the functional assay(seebelow) were pooled, lyophilized,and subjected to gel filtration chromatographyin PBS on a TSK 3000SW preparative column (LKB-Pharmacia,Gaithersburg, MD). The purity of the ehted peptides was checkedby SDS-PAGEaccordingto Laemmli (38), using a mini-gel system (Bio-Rad Laboratories, Richmond, CA). Partial amino acid sequences were determined on a protein sequencer (475 A; Applied Biosystems, Inc., Foster City, CA). Antigen-specific11.,2ReleaseAssay. The T cell hybridoma (5 x 104 cells/well) was tested with a titration of collagens or CII CNBr peptides in the presence of DBA/1 or SWR/J spleen cells (5 x 10s cells/well) in flat-bottomed 96-well plates (Costar, Cambridge, MA). After 20-24 h of incubation, cell-free supernatants were collectedand tested for I1,2 content using the CTL1,2 proliferation assay. CTL1,2 were plated at 5 x 103/well, and after 2024 h of culture were pulsed for an additional 6 h with 1/~Ci/well of [3H]thymidine (Amersham International, Amersham, Bucks UK). Incorporated radioactivity was measured with a liquid scintillation B counter (Betaplate; LKB-Wallac,Turku, Finland) and is expressed as mean cpm of triplicate cultures. Induction and Evaluation of CIA. Mice (8-10 wk) were immunized by intradermal injection in the back at the base of the tail with 200/~g of bovine or chicken CII in CFA (Difco Laboratories, Detroit, Michigan). An intraperitoneal injection of Mycobacterium bovis Bacillus Calmette-Gu6rin (BCG, ,,o0.8-2.6 x 106 cells/mouse; Schweiz. Serum- & Impfmstitut, Bern, Switzerland) was also given to the mice at the same time. 21 d after the immunization, mice were boosted intraperitoneally with 200 #g of CII together with BCG. Arthritis was monitored daily from day 21 until day 35, weekly during the second month, and then every 2 wk thereafter. Mice were monitored for up to 3 mo, and in the case ofSWR fl tg mice, for up to 5 mo after immunization. Dates of onset of disease were recorded for individual mice. CIA was diagnosed if redness and swelling of fore- and hindpaws were observed. Severityof arthritis was evaluatedfor each paw and graded as following: grade 1, redness and swelling; grade 2, deformity; and grade 3, ankylosis.The scores for each paw were added to give an index ranging from 0 to 12 per mouse. None of the mice ever developed CIA after CFA/BCG injection. Statistical analysis was done using the X2 test (for comparing CIA incidence) and the Mann-Whitney test (for comparing day of onset and severity of CIA). TCR 3 Gene Constructand Generation of SWR ~ tg Mice. The productively rearranged TCR B gene has been isolated by PCR amplification from genomic DNA of the anti-CII T cell clone. Primers used are complementary to sequences located in the 5' flanking region of the VB12 leader sequence containing an additional SalI cloning site (5' primer), and in the intron between JB1.2 and JB1.30' primer), allowing the amplificationof a DNA region containing a BamHI site at the 3' ofJB1.2, pCR was performed for 35 cycles (94~ 40 s; 65~ 50 s; and 72~ 60 s) followed by 5 rain at 72~ using Taq DNA polymerase(Perkin-ElmerCetus, Emeryville, CA). The amplifiedfragment of ~900 bp was inserted into a plasmid vector containing TCR B enhancer and VB8.3 pro-
Collagen-inducedArthritis in TCP. 13TransgenicMice
moter sequences (C. E Gray, Hoffmann-La Roche, Basel, Switzerland). A 6.4-kb fragment spanning the region from the BamHI site located in the J/31 locus to the HindlII site at the end of the C/31 untranslated region, was further ligated into the construct at the corresponding BamHI site (see Fig. 2). This provides the entire C/31 region used by the K-102 T cell clone. After DNA sequencing to control possible PCK and ligation artifacts, the/3 construct was excised from the prokaryotic vector (pBluescript II KS- ; Stratagene, La Jolla, CA) by digestion with BssHII, separated on agarose gd, and purified by phenol/chloroform extractions followed by ethanol precipitation. Transgenic mice were generated by microinjection of the purified DNA into SWR/J fertilized oocytes that were reimplanted into the oviducts of MOP,O/Ibm foster mothers as described (39). The integration of the transgene was analyzed both by PCR and Southern blot using DNA isolated from tail biopsies of 2-3-wk-old mice. The transcription of the fulllength tg B RNA was demonstrated by Northern blot using total spleen RNA. AntibodiesandFlowCytometry. The following mAb were used: anti-CD3 e biotin-conjugated (500A2; PharMingen, San Diego, CA) or FITC-conjugated (145-2Cll; Boehringer Mannheim, Mannhelm, Germany); anti-CD4 PE-conjugated (GK1.5; Becton Dickinson & Co., Mountain View, CA); anti-CD8 FITC-conjugated (53-6.7; Becton Dickinson & Co.); anti-Thy 1.2 FITC-conjugated (30-H12; Becton Dickinson & Co.); anti-VB2 (B20.6.5, kindly provided by B. Malissen, MarseiUe-Luminy, France); anti-V~3 (KJ25a; J. Kappler, Howard Hughes Medical Institute, Denver, CO); anti-VB4 (KT4-3; K. Tomonari, Medical Research Coundl, Harrow, U.K.); anti-V~6 (44-22-1; H. Hengartner, University Hospital, Ziirich, Switzerland); anti-V~8.1, 8.2, and 8.3 (U. Staerz, National Jewish Center, Denver, CO); anti-VBll (KT11; K. Tomonari); and anti-V~17a (KJ23a, J. Kappler) were all used as cell culture supernatants. The MR11-1 hybridoma secreting anti-V/312 mAb was produced by O. Kanagawa (Washington University, St. Louis, MO), and obtained through the courtesy of H. R. MacDonald (Ludwig Institute for Cancer Research, Lausanne, Switzerland). The antibody was purified from the cell culture supernatant and biotinylated by standard procedures using N-bydroxysuccinimidobiotin (Sigma Chemical Co.). Second step reagents were: sheep anti-mouse Ig-FITC, and sheep anti-rat Ig-FITC (Silenus, Hawthorn, Australia); goat anti-mouse Ig-PE, and goat anti-rat Ig-PE (Southern Biotechnology Associates, Birmingham, AL); and streptavidin-FITC or streptavidin-conjugated tandem label of PE/Texas Red (Southern Biotechnology Associates). All incubations and washings were done at 4~ in PBS, 2% FCS, 0.02% NAN3. 20,000 or 50,000 viable cells were analyzed by FACScan| (Becton Dickinson & Co.). Dead cells were excluded from the analysis using forward and side scatter parameters and also using propidium iodide when possible. Serum Anti-CII Antibody Levels. ELISA was performed as follows: fiat-bottomed Maxi Sorp 96-well Immuno Plates (Nunc, Roskilde, Denmark) were coated overnight at 4"C with 50/~l/well of either CII or purified protein derivative from Mycobacteriumtuberculosis(PPD, Statens Seruminstitut, Copenhagen, Denmark), BSA (Sigma Chemical Co.), or goat anti-mouse Ig (Southern Biotechnology Associates), all at 10/zg/ml. Plates were saturated with 1% BSA in PBS, and after washing, sera were added at various dilutions. After five washes with PBS, 0.05% Tween 20, ~ galactosidasehbeled antibodies (anti-mouse Ig, IgM, IgG1, IgG2a, IgG2b, IgG3, Southern Biotechnology Associates)were added and incubated overnight at 4~ Positive wells were revealed with O-nitrophenyl-~n-galactopyranoside (Sigma Chemical Co.) and ODium was determined using a microplate reader (Anthos Labtec, Basel, Switzerland). 383
Mori et al.
Results
The T Cell Hybridoma BL17 Recognizes the CBlI Fragment of CII. A T cdl hybridoma, BL17, was derived from the fusion of the arthritogenic anti-CII T cell done K-102 and the BW-1100.129.237 c~-~- T cell lymphoma (see Materials and Methods). This hybridoma expresses high levels of CD3 and CD4. Stimulation by immobilized anti-c~/~ TCR or antiCD3 antibodies leads to Ib2 release and apoptosis (data not shown). Like the parental clone, the hybridoma produces Ib2 in response to bovine, chicken, human, and mouse CII presented by DBA/1 APC (data not shown). To further characterize the antigen specificity, bovine and chicken CII were digested with CNBr, and the peptides were fractionated by ion exchange and gel filtration chromatography. The fractions were tested for their ability to induce Ib2 release in the presence of APC. Stimulatory activity could be attributed to a peptide migrating with an apparent molecular mass of about 36 kD on SDS-PAGE (data not shown). Partial amino acid sequences of the stimuhtory peptides from both bovine and chicken CII showed that they were identical with the previously described CBll peptide (40), a 279-amino acid-long peptide that is the main CII fragment recognized by both antibodies and T ceils in mice with CIA.
$ WR/JSpleen Cells Can Processand PresentCII and its CB11 Fragment to the T Cell Hybridoma. S W R / J mice are not susceptible to CIA induction, perhaps because of the deletion of some V~ gene segments. An alternative explanation could be that although they carry susceptibility-associated H-2 genes, their APC might be unable to present arthritogenic CII epitopes. We have excluded this possibility by showing that spleen cells from S W R / J mice are able to process and present either bovine, chicken and human CII (Fig. 1 a), or bovine and chicken CBll (Fig. 1 b) to the BL17 hybridoma. Thus, the resistance of S W R / J to CIA is not due to a defect in antigen presentation. Generation of TCR B TransgenicMice. We have attempted to sensitize the S W R mice to CIA by introducing a transgene encoding the BL17 TCR ~ chain. This T C R B chain was shown to be encoded by V/512-D~l.l-J~l.1 and C~/1 (L. Mori, unpublished results). The productively rearranged
80
~
10
40 ZO
.01
.1
1
I0
1~
IOOQ .OOG1.001 .01
.1
1
10
pglml
Figure 1. Antigenpresentationby SWR/J spleencells. BL17cellswere stimulated by CII or its CBll fragment in the presenceof SWR/J APC. The levelsoflb2 secretedweremeasuredusingthe CTLL-2assay.(a) Shows the stimulation with bovine CII (11), human CII (0), and chicken CII (A); (b) with bovine CB11 ([~) and chicken CB11 (A). Results are expressed as mean cpm of [3H]thymidineincorporation of triplicate cultures.
LVI312DI31.1J]31.1 5'
El3
PVI~8"31 ~
131
a SWR/J
Figure 2. TCR ~ construct used for the generation of transgenic mice. The construct is composed of a PCR-amplified fragment that contains the leader sequence of VB12 (L), V~12, DBI.1, and JBI.1 (nucleotide sequence is available upon request). This fragment is inserted 3' to the/3 locus enhancer (E/3) and promoter (P) of V/iS.3, and 5' to the entire CB1 lOCUS.
~_
TCK B gene was isolated by PCR amplification of genomic DNA isolated from the anti-CII T cell clone K-102, and inserted into a plasmid vector containing the TCR/3 enhancer and VBS.3 promoter sequences. A fragment encompassing the entire C/51 locus was further inserted at the 3' end of the V/5 gene (Fig. 2). This TCK/5 construct, free from v e c t o r sequences, was used for the generation of tg mice. Two tg mouse lines were obtained in the SWK/J strain. All experiments described in this study were carried out with one of these lines (SWR-/SL tg) which integrated about 10 copies of the transgene into the genome and expressed very high levels of tg/5 chain mRNA in the spleen (data not shown).
Expression of the TCR /8 Transgenein the Absence of Endogenous B Genes. The expression of the TCR/5 transgene was tested using a mAb specific for the mouse V/512 polypeptide. Thymocytes and lymph node cells from tg mice and non-tg littermates were stained with mAb specific for CD4, CD8, and V/512, and evaluated by three-color FACS| analysis. In the thymus, V/512+ calls were undetectable in SWK mice, as expected, but were present in SWR-BL tg animals (Fig. 3, a and d). The number of V/512+ and CD3 * ceils
r
b 36
3
SWR/J 5d
e
~s
~
f
20
~3
[ .*
'" 4
SWR [3L tg
V~12
CD3
CD8
Figure 3. Thymocyte surface expression of the transgenic/3 chain in SWR-BL tg mice. Thymocytcs from SWK-BL tg and non-tg, sex-matched, littermate mice were stained with anti-CD8 (FITC-conjugated), anti-CD4 (PE-conjugated) and anti-V/~12 or anti-CD3 (biotin-conjugated), followed by labeling with streptavidin-tandem (see Materials and Methods). (a and d) Histograms of the Vf112 staining; (b and e) Histograms of the CD3 staining; (c andf ) Contour plots for CD4 and CD8 stainings ofceUs gated as CD3 + . Numbers represent the percentage of cells in each region. The difference in staining of the TCR. k'~ population, using anti-Vf112 or antiCD3, has to be attributed to different ai~nities of the two mAb. 384
iI
e L~ r~
f
t
d SWR JSLtg
V~312
a
c
3' lKb
Tg ~
b
CD3
CD8
Figure 4. Lymph node T cell surface expression of the transgenic/~ chain in SWR-BL tg mice. Total lymph node cells from SWR-BL tg and nontg, sex-matched, littermate mice were analyzed by three-color staining as in Fig. 3. (a and d) Histograms of the V~12 staining; (b and e) Histograms of the CD3 staining; (c and f ) Contour plots for CD4 and CD8 stainings of cells gated as CD3 +. Numbers represent the percentage of cells in each region.
was similar, thus indicating that every T cell expresses the transgene. Although the total number of CD3 + thymocytes was similar in normal and tg animals, there were differences in the percentages of mature cells. Tg mice had a lower number of CD3 hi thymocytes (3% vs 12% in normal animals) and this was compensated by a higher number of immature, CDY ~ cells (49% vs 45%; Fig. 3, b and e) and CD3+CD4+CD8 + cells (63% vs 53%; Fig. 3, c and f). In lymph nodes of tg mice all CD3 + cells were V/512+ and were reduced in number (52% vs 74% of total lymph node cells; Fig. 4, a, b, d, and e), but the ratio of CD4 to CD8 cells was as in normal mice (Fig. 4, c and f). We also used other VO-specific mAb (anti-VB 2, 3, 4, 6, and 17a) to show the lack of expression of endogenous/3 genes. Indeed, none of these reagents stained either thymocytes or lymph node cells in SWR-/SL tg animals (data not shown). These results show that the tg/~ chain is expressed on the surface of T cells and that it prevents the expression of the endogenous TCR/3 chains. Furthermore, expression of this chain also affects the number of mature cells, without altering the relative percentages of CD4 + and CD8 + subsets.
SWR-~L tg Mice Mount Strong Anti-CII Antibody Responses. The antibody response of mice to CII is strictly T cell dependent (41). SWK/3 tg and non-tg littermates, as weU as DBA/1 mice were immunized with bovine CII in CFA. At different time points after immunization, mice were bled and the anti-CII antibodies were analyzed by ELISA. It is interesting that the SWR-BL tg mice produce more antiCII antibodies (day 7 titer, 8; day 14 titer, 50; and day 21 titer, 400) than DBA/1 mice (day 14 titer, 15, and day 21 titer, 200) and SWR non-tg littermates (day 14 titer, 3, and day 21 titer, 25; Fig. 5). Furthermore, their response developed earlier as compared with the other groups. Already at day 7 after immunization, SWR-/SL tg mice had detectable titers of anti-CII antibodies of the 3/1, 3'2b, and 3'2a isotypes. The anti-CII antibody titer in SWR-/SL tg mice was always
Collagen-induced Arthritis in TCR/3 Transgenic Mice
1.2
Table 1. SWR-~L tg Mice are Resistant to CIA
Day 7
Mice*
0.8
Arthritic/Total
%
0.4
Experiment 1 0.0
0.8
~2 0
0.4"
0.0
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
1.2 0.8'
0.4'
0.0
.
10 0 101
.
.
.
.
10 2 10 3 10 4 10 5
1 : serum dilution
Figure 5. Increasein the total anti-CII antibodytiter in the primary response of SWK-BL tg mice. Titration of the total antibovineCII Ig presentin the serum of mice immunized with bovine CII in CFA (opensymbols)or only injectedwith CFA(cbsedsymbols) on differentdaysafter immunization: SWR-/~L tg ([], I); SWR/J (A, a); DBA/1 (O, O). Each curve represents the values obtained from the pooled sera of three mice. Results are expressed in OD~nm. Control ELISAwith BSA-coated plates gave background values ~0.172.
higher than in S W R non-tg littciraates for all antibody dasses and at all time points (data not shown). Thus, in the SWR-BL tg mice there seems to be an early switch to all Ig isotypes and in particular to ~/2a that predominates in the anti-CII antibody response during CIA (41). This finding, together with the fact that the antibody titer is higher in the tg mice, suggests that these mice may have an increased number of CII-specific Th cells. SWR-~L tg Mice Are Resistant to CIA. SWR-/~L tg mice were tested for the development of CIA, to determine whether the expression ofa Vf112 bearing TCR/3 chain on all T cells, and the increased production of anti-CII antibodies observed in these mice were sufficient to make them susceptible to the disease. A total of 31 SWIL-/~ tg mice and 27 S W R non-tg littermates were immunized with bovine CII in CFA in three separate experiments. 21 d after the immunization, mice were boosted with bovine CII and analyzed for the development
Table
Incidence
2.
Experiment 2
Experiment 3
SWK-BL tg SWR/J DBA/1 SWIL-/~Ltg SWR/J DBA/1 SWR-/3L tg SWR/J DBA/1
0/12 0/14 7/11 0/13 0/8 6/7 0/6 0/5 4/5
0 0 64 0 0 86 0 0 80
* Male and female mice were used.
of arthritis. As shown in Table 1, none of the S W K developed clinical signs of CIA regardless of the presence of the transgene, even 5 mo after the first injection. In contrast, 64-86% of the DBA/1 mice showed clinical signs of CIA, starting from day 21 to day 33 after CII immunization. Expression of the tg ~ Chain in Susceptible Mice Leads to In. creased Incidence and Severity of CIA. Susceptibility to CIA is a dominant trait since (DBA/1 x SWK) F1 mice develop CIA (10, 11). Both tg and non-tg F1 littermates, derived from crossing S W K mice heterozygous for the/~ tg with DBA/1, were immunized with CII and observed for the development of CIA. Arthritis developed in 38% of (DBA/1 x SWK) F1 mice and in 89% of the F1/3L tg mice in a total of three independent experiments (Table 2). The disease induced in F1 BL tg mice was more severe (median arthritic score 8 vs 4) and developed 2 d earlier than in F1 nontg littermates. V/312 was expressed in the tg mice by all T cells present in lymph nodes draining the affected limbs, as well as in the corresponding lymph nodes of the tg mice without clinical CIA (Fig. 6). FACS| analysis using other V/3-specific mAb (anti-V/3 2, 3, 4, 6, 8, 11, and 17a), showed that none of the F1/~L tg animals also expressed endogenous/3 genes (data not shown).
CIA in (DBA/1 x SWR-ffL tg) F~ Mice
Mice
Arthritic/Total
Incidence*
Day of onset**
Severity* **
26 (21-29) 28 (28-36)
8 (12-3)
% F1/3L tg littermates Ft non-tg littermates
25/28 8/21
* p
